structural integrity of dental crowns878377/fulltext01.pdf · there is currently no standardized...
TRANSCRIPT
DEGREE PROJECT, IN ,ENGINEERING MECHANICS - SOLID MECHANICSSECOND LEVEL
STOCKHOLM, SWEDEN 2014
Structural integrity of dental crowns
NATHALIE HAMSUND
KTH ROYAL INSTITUTE OF TECHNOLOGY
SCHOOL OF ENGINEERING SCIENCES
Acknowledgements This report is a Master thesis project in Solid Mechanics at the Royal Institute of Technology (KTH).
The work was performed at the Mechanical Engineering department at Queen´s University in Belfast,
Northern Ireland during the spring 2014. Doctor Elaheh Ghassemieh was the supervisor at Queen´s
University and Professor Bo Alfredsson was the examiner at KTH.
I would like to take the opportunity to express my deepest gratitude to my supervisor Doctor Elaheh
Ghassemieh for her great support through the project and continuous feedback. Along the way I
have also received valuable advice from Professor Bo Alfredsson and Associate Professor Artem
Kulachenko.
Summary There is currently no standardized method for testing dentures made of dental ceramic materials.
This creates problems for comparisons between studies concerning strength in materials and
geometry. The geometry of the maxillary first premolar was used for the analysis. The modeled
crown has two underlying layers, a thin cement layer closest to the crown that attaches the crown to
dentin. Simplifications are made so that dentin geometry is modeled in CAD as a cuboid and the
cement layer above follow its geometry. For the actual analysis, a FE-analysis has been made of
different loading positions and areas to clearly show how important it is to be consistent in the
selection of these. The load has been placed on two surfaces with 150 N each to resemble a real
occlusal maximum load. The standard area used in other analysis was set to be 5.5 mm2 on each
cusp. A comparative analysis has been made of various ceramic materials to see what impact it has
on the cement layer strength. Similarly the cement material properties were modified to see how it
affects the ceramic crowns strength. In real applications, one usually tries to achieve as thin cement
layer as possible but here the analysis is of how much influence the cement layer thickness has on
the overall crown strength. Two cylindrically shaped simplified models were created to see how
much the anatomical geometry effects on the maximum stresses. One was modeled with a flat top
and the other with an angled top.
The load placement and the magnitude of its area turned out to be very important for the resulting
maximum stresses. Surface area ranged from 0.07 - 26 mm2, and the resulting von Mises stresses for
these ranged from 1120 to 34 MPa. Generally speaking, varying cement materials available on the
market does not give huge impact on ceramic bearing stress. Panavia F 2.0 was found to be the
strongest cementing layer that caused the highest failure load of the existing materials on the
market. Super Bond B&D gave the lowest failure load. However, you could see that it was a bit more
important to be careful for the crown material. The made up ceramic materials (Hypothetical
Ceramic and Experimental Ceramic) proved to cause relatively close stresses for the cement to fail. It
was also clear that a thicker cement layer reduces the expected failure load and the structure
becomes weaker. The simplified models seemed to give a substantial exaggeration of strength. The
expected failure load was nearly twice as large for these (slightly lower for the angled model) than
for the anatomical model.
When comparing with the test data from another study it proved that it does not match with the
simulations. In the study different cementing materials had been tested and Super Bond B&D turned
out to be the strongest material. With this in hand it can be said that a stable test setup with
consistently identical geometries is necessary in order to evaluate these kinds of structures.
Sammanfattning Idag finns ingen standardiserad provninvgsmetod för tandproteser gjorda i odontologiska
keramikmaterial. Det skapar problem vid jämförelser mellan studier som rör hållfasthet med
avseende på material och geometri. I det här examensarbetet har en geometri från övre käkbenets
första premolar analyserats. Den modellerade kronan har två underliggande lager, ett tunt
cementlager närmast kronan som fäster kronan till dentin. Förenklingar är gjorda så att dentinets
geometri modellerats i CAD som ett rätblock och cementlagret ovan följer dess geometri. Vid själva
analysen har man gjort FE-analys av olika lastpositioner och areor för att tydligt visa hur viktigt det är
att vara konsekvent vid valet av dessa. Lasten har lagts på två ytor med 150 N på vardera kusp för att
likna ett verkligt betts högsta lastkraft. Som standardfall att använda för andra analyser valdes en
area på 5.5 mm2 på vardera kusp. En jämförande analys har gjorts av olika ceramiska material för att
se vilken inverkan det har på cementlagrets hållfasthet. På samma sätt ändrades cement materialets
egenskaper för att se hur det inverkar på kermaikkronans hållfasthet. I verkliga applikationer
försöker man oftast nå ett så tunnt cementlager som möjligt, här har en analys gjorts av hur stor
inverkan på hållfastheten som cementlagret har. Två cylindriskt formade förenklade modeller
skapades för att se den anatomiska geometrins inverkan på spänningsbilden. En modellerades med
platt topp och den andra med vinklad topp.
Lastens placering och arean den är utbredd på visade sig vara väldigt viktigt för de resulterande
maximala spänningarna. Ytans area varierade mellan 0.07 – 26 mm2 och resulterande von Mises
spänningarna för dessa varierade 1120 – 34 MPa. Generellt sett hade olika cementmaterial som finns
på marknaden inte jättestor inverkan på keramiklagrets spänningsbild. Panavia F 2.0 visade sig ändå
vara det starkaste cementeringslagret som orsakade högst brottlast av de befintliga materialen på
marknaden. Superbond B&D gav lägst resultat. Däremot kunde man se att det var lite viktigare att
vara noggrann med kronans materialval. De påhittade keramikmaterialens (Hypothetical Ceramic och
Experimental Ceramic) spänningar visade sig komma realtivt nära cementlagrets brottgräns. Det
visade sig också tydligt att för ett tjockare cement lager minskar den förväntade brottlasten och
strukturen blir svagare. De förenklade modellerna tycktes ge en kraftig överdrift i hållfasthet. Den
förväntade brottlasten var nästan dubbelt så stor för dessa (något lägre för den vinklade modellen)
än för den anatomiska modellen.
Vid jämförelse med testdata från en annan studie visade de sig inte stämma överens med
simuleringarna särskilt bra. Man hade testat olika cementeringsmaterial och där visade sig
Superbond B&D vara det starkaste materialet. Utifrån det kan man säga att en stabil
provningsuppställning med konsekvent identiska geometrier är nödvändigt för att kunna evaluera
sådana här strukturer.
Table of Contents Introduction ............................................................................................................................................. 1
Project aim .......................................................................................................................................... 1
Background .......................................................................................................................................... 1
Methods .................................................................................................................................................. 2
Model from STL file ............................................................................................................................. 2
Creating the solid geometry for anatomic model ........................................................................... 3
Preparing the model for analysis 142 high 16-63-140 ................................................................... 4
Loading cases ................................................................................................................................... 5
Simplified models ................................................................................................................................ 6
Materials .............................................................................................................................................. 7
Calculating expected failure load ........................................................................................................ 8
Results and Discussion ........................................................................................................................ 9
Load application .................................................................................................................................. 9
Stresses in Ceramic when varying Cement thickness........................................................................ 10
Resulting stresses for simplified models ........................................................................................... 11
Different materials in Ceramic and Cement ...................................................................................... 12
Varying Ceramic material .............................................................................................................. 12
Varying Cement material ............................................................................................................... 13
Conclusions ............................................................................................................................................ 15
References ............................................................................................................................................. 16
1
Introduction This study is a Master thesis on structural integrity of dental crowns. The crown used in the
numerical analysis is a maxillary first premolar. That means the first tooth on the upper jaw with two
tips. The ceramic crown is attached to dentin with a thin cement layer. In this study the dentin core
material is approximated to be the only inner material, excluding the pulp. The setup is shown in
Figure 2 in the Methods section.
Project aim This study was performed to confirm that the chosen loading position and area is crucial to the stress
results and expected force at failure. It will also show how sensitive this kind of structure is to flaws
in geometry by comparing simulation with test results.
Background In earlier studies the fracture strength has been measured in testing of flat vs anatomic design of a
dental crown. The flat design has showed to possess a slight higher strength [1]. Although a flat
design might not be enough to use for modelling a crown, evaluation of for example material can still
be extracted from such a test.
Earlier studies say that the highest load a human tooth is subjected to is about 150 - 665 N in total
[2]. The load is distributed differently depending on the tooth position and characteristic.
When making physical testing it is hard to reconstruct the occlusal area that receives most of the
load. For a maxillary first premolar tooth the load area is distributed on the inner cusps. The actual
area has not been well documented in literature but an average of a wear facet on a tooth is believed
to be of 0.5 - 3 mm in diameter [2]. If this area is considered to be perfectly circular the area would
be 0.2 – 7 mm2. The reconstruction is often made by loading the tooth at its centre with a spherical
indenter with a small diameter, 4 - 6 mm. The material of the ball is not consistent in various studies,
sometimes it has been of steel [3] and for another study it has been of tungsten carbide [4]. The
indentation area that it creates can be estimated by Hert’z theory and gives approximately 0.01 mm2.
According to literature a steel ball would have to be of a much larger diameter to be representing a
real case scenario, about 40 mm to 1 m [2].
The cement thickness in real applications varies between 0.03 – 0.1 mm [2]. According to material
specification of cement materials the minimum film thickness with bonding properties is 0.16 mm
[5].
Dentin strength and elasticity differs a lot between individual persons. It is very much dependent on
the person’s age, gender and general oral health [6]. Also the material properties within the dentin
differ through the layer being stronger closer to the pulp at the centre [7]. In this study the dentin
material properties are chosen from an existing article making similar evaluations on a premolar
model in FE analysis [8].
2
Methods Three different models were constructed throughout the project. First the anatomical one from a STL
file provided within the projectand thereafter two very simplified axisymmetric ones. The first one
with a flat top surface and the second with an angled top, the thickness of the ceramic layer
decreases closer to the centre. The comparisons were mainly made by looking at the maximum
principal stress for compression and maximum von Mises stress for bending. The material properties
used for the original setup are shown in Table 1 [2] [8]. All models were created as linearly elastic and
isotropic for convenience [9].
Table 1 The material properties used for the original standard setup.
Material Young´s Modulus
[GPa] Poisson´s ratio
Flexural strength [MPa]
Compressive strength [MPa]
Dentin (Average) 18.6 0.31 212.9 297 Cement (RelyX luting plus) 4 0.35 34 135 Ceramic (IPS e. max Press) 94.4 0.23 374.4 360
Model from STL file The STL file only contains triangles building the geometry and information of the relation between
them. It does not contain any information of the actual magnitude of the geometry. Three samples of
dental crowns were photographed in a micro lab to provide a high accuracy for measurement to
decide an average geometry shown in Figure 1.
Figure 1 Pictures from micro lab that were used for determining an average geometry used as standard in all models.
The STL file contained triangles building a maxillary premolar crown. It had to be modified in order to
make simulations with additionally two layers. The layer closest to the ceramic crown layer was a
thin cement layer. The rest was filled with a dentin material. This is more clearly defined in Figure 2.
Figure 2 Cut view showing the concept of the original setup model.
3
Creating the solid geometry for anatomic model
A STL file is built up by triangles creating a surface. To convert it into a solid part it was imported into
Meshlab [10] to reduce number of areas by performing Quadratic Edge Collapse Decimation as the
original file contained an exceeding number of areas for importing into any of the software available
with a student license. The STL file with reduced number of areas was imported into Solidworks [11].
There it was converted into a solid body and exported as an IGES file. The next step was to import
the solid body into Solid Edge [12] and modify it into the model used for FE simulations in Ansys
Workbench [13].
The Quadratic Edge Collapse Decimation performed in Meshlab [10]resulted in a dental crown
model with 2000 surfaces. The STL file does not contain any information about dimensions or
magnitude. So when the model was imported to Solidworks [11] the selection of dimensioning
system was set to be meters. This created a very large model (approximately 3 m x 4 m in the
transactional base area instead of the selected average of 6 mm x 10 mm) but the magnitude was
consistent through all used software. When modifying the model in Solid Edge [12] a number of
planes were created underneath the ceramic crown in order to first fill it with material by sketching
fitting geometries and using the Extrude command. An additional sketch was made at the bottom of
the tooth for creating the geometry that had to be cut-out in order to fill it with the other materials
as shown in Figure 3.
Figure 3 Ceramic layer modified from STL file.
The cement and dentin layer were created as standalone parts. An assembly was constructed from
the three layers mentioned and exported to Ansys Workbench [13] in a parasolid format. The cement
layer was varied in order to see its importance for the resulting maximum stresses. The thicknesses
tested were 0.05, 0.1, 0.2, 0.3, 0.4 mm. In order to do this a CAD model had to be created for each
setup. The midline of the cement thickness was kept constant and is shown in Figure 4.
4
Figure 4 Geometry for midlines in cement layer.
Preparing the model for analysis
In order for the solution to be reliable a mesh convergence had to be made. To do this a coarse mesh
was set to the model and the analysis was run. The element size of the mesh was then gradually
refined and an analysis was run between every refinement. By doing this it could be seen to what
stresses the solution was converging. If the mesh would not have converged it would probably mean
that there are some sharp edges creating high stress concentrations. For this case convergence was
found for a uniform mesh with a gradual refinement in the edges where the cusps meet as seen in
Figure 5. The refinement at the sharp edges is done by splitting every tetrahedron into nine elements
improving the accuracy of the structure locally. The elements closest to these were split into
additionally four elements for every element. This was to create a gradual refinement close to peak
geometry.
Figure 5 A uniform mesh through all layers with refinement at sharp edges.
The elements used were tetrahedrons (SOLID187 in Ansys) as those are most convenient for a
complex geometry. The total number of elements was 121 664. Through the cement thickness there
was one tetrahedron element. Every element has ten nodes which Ansys uses to solve in the analysis.
For all thicknesses higher than the standard setup (>0.1 mm) there were two to three elements
5
through the cement layer thickness. Between the different material layers a bonded contact was
defined. It means that when contact occurs between the bodies the surfaces becomes tied together
and cannot slide with respect to each other. This seems right as the cement material in the middle
should have bonding properties.
Loading cases
The model was loaded in three different ways. The first method was according to indentation loading
in two targets at the cusps creating two-point loading at the two highest points of the ceramic layer
shown in Figure 6. A distributed load of 150 N was attached to a small area calculated from Hertz
theory with generalization showed in Figure 7. From literature it was found that for practical testing
the indentation often is a ball, many times in steel with some differences in radius. Here the
properties of the steel ball was set as R = 2 mm, E = 200 GPa, ν = 0.3 [1].
Figure 6 Loading position when loading at ceramic tips.
Figure 7 Concept of ball indentation into flat surface used for determining indentation area.
An equivalent elastic modulus was determined according to (1).
(1)
Further the indentation area was calculated by equation (2)-(4).
(
√ )
⁄
(2)
6
√ (3)
(4)
The value of indentation area on the real case, , calculated from equation (4) was used on both
tips to distribute the load over it. That area was doubled and increased a number of times to see
what an increasing area at the tips would do to the maximum stresses. The loading areas tested were
0.07, 0.14, 0.28, 0.42 and 0.55 mm2 for each tip. As mentioned above all parameters used on the STL
model had to be magnified. The general reference areas were set from the transactional bottom area
of the crown as Aref = 60 mm2. The areas of the model used in simulations are denoted with zero. The
transactional bottom area of the model was A0 = 13.3 m2. The relation described in equation (5) gave
a value to use for the indentation area, , in the model.
(5)
The second way of applying load was to distribute it on the cusps accordingly shown in Figure 8. This
is a more probable loading case and more like a real life occlusal area [2]. Four different areas of this
type were tested, 5, 5.5, 7.6 and 13 mm2 on each cusp. The second area, 5.5 mm2 and this loading
type was used as the standard setup configuration for all analyses.
Figure 8 The Cusp loading case which is similar to real-life occlusal areas.
The last way of loading is the area attained in experiments described in literature [1] and [9]. This
loading area is in total 25 mm2, about 12.5 mm2 at each top creating a constant pressure and is
shown in Figure 9.
Figure 9 The tip loading case when the area that occurs can be similar to compressive testing.
Simplified models Two simplified models of dental crowns were created in CAD [12] and analysed in FE software [13].
The setup geometries are shown in Figure 10, one with flat top and one with angled top. The
cylindrical models were set to have a corresponding geometry as in the anatomic model. The base
7
area of the anatomic model is the same as in the cylindrical base area. In the angled configuration
the angle was decided throughout that the ceramic layer should not get thinner than 1 mm as the
ceramic in real restorations generally are made with thickness 1 – 2 mm [2]. The load was applied to
a total area of 10 mm2, about 5 mm2 on each cusp, in order to make the loading area similar to the
standard setup configuration. In the angled model the angle in the middle is reduced to avoid stress
singularities by the peak geometry shown in Figure 11.
Figure 10 Geometries of the cylindrical setups.
Figure 11 Refinement in middle of angled cylindrical model
Materials The materials influence on the stress distribution of the crown was analysed. Both the cement and
ceramic were varied in its properties. When varying the ceramic, seen in Table 2, it was seen how the
maximum stresses in the cement layer varied. Accordingly the maximum stresses in the ceramic were
analysed when the cement layer material properties were varied, seen in Table 3.
8
Table 2 The ceramics used for analysis.
Ceramic Material E-moduli [GPa] Poisson's ratio Hypothetical Ceramic (low elastic modulus)
20 0.24
IPS Empress 65 0.23 Average Ceramic 69 0.25 IPS e.max Press 94.4 0.23 Experimental Ceramic 103 0.24
Table 3 The cements used for analysis.
Cement Material E-moduli [GPa] Poisson’s ratio Super bond C&B, self-cure dental adhesive resin cement
1.8 0.25
RelyX luting plus 4 0.35 Average Cement 5 0.3 Rely X Unicem, dual-polymerized self- adhesive universal resin cement
8 0.33
Panavia F 2.0, dual-polymerized phosphate- modified resin cement
18.3 0.3
Hypothetical, extreme high value material 40 0.35
Calculating expected failure load When calculating the expected load for failure it was assumed that the load-to-failure value would be
linear as this is a linear elastic model explained in (6).
(6)
The differences in expected load to failure are due to von Mises and Rankine theory. Rankine failure
criterion only look at the maximum Principal stress, the most positive or most negative principal
stress and use the compressive or tensile strength respectively as a reference as seen in (7).
| |
(7)
When using von Mises failure criterion the value used in the scale factor takes into account for all
principal stresses. In the 3D case this means the 1st, 2nd and 3rd Principal stress. Then this value is
compared to the flexural strength according to (8). The flexural strength of a material is determined
by a three point bending test is used as measurement for bending strength in brittle materials.
√
[
]
(8)
The relations of (7) and (8) shows that the different failure criterions have little to do with each
other. They do not use the same reference values and the output value for comparison depends a lot
on the Principal stresses relation in magnitude.
9
Results and Discussion For all analysis cases in Ansys the solution became non-linear. This is expected for a model with contact elements [14]. The expected failure load for the original setup is seen in Table 4. The original setup means that the applied load is 150 N at 5.5 mm2 on each cusp and cement thickness of 0.1 mm. The ceramic material used was IPS e.max Press and the cement material used was RelyX luting plus. Table 4 Expected failure load according to different strengths.
Expected failure load according to von Mises [N]
Expected failure load according to Rankine in
compression [N]
Expected failure load according to Rankine in
tension [N] 977 759 1594
According to these simulations the dental crown seems to be most sensitive to compression and least to tension. The expected failure loads seem reliable when comparing to another study where testing has been made of ceramic crowns of IPS e.max [15]. The failure load in that study was mainly between 800 – 850 N. A specification of failure mode was not reported as the samples generally were totally demolished.
Load application The expected failure load showed to be very much dependent on the loading area. This can be seen
from the diagram in Figure 12. The flexural strength for the ceramic is seen by a line in the diagram.
For a very small loading area presented on the right hand side the flexural strength seems to cause a
problem. According to the flexural strength, failure in ceramic has occurred for less than half the load
applied here when loading at the tips on 0.07 mm2.
Figure 12 Diagram showing the maximum von Mises stress according to loading position and area. The flexural strength for the ceramic layer (IPS e.max Press) is marked out with a line.
The simulation made for compression on tips has a slightly smaller area than the largest load area on
the cusps. The resulting maximum von Mises stress seems to be higher when loading at the cusps
than when loading at the tips. This implies that a compressive test by loading at the tips can create a
10
risky underestimation of the actual maximum stress that occurs in a real case where loading mainly
occurs at the cusps.
Stresses in Ceramic when varying Cement thickness The results from changing the cement thickness are presented in Table 5. It is seen that the stresses
increase for an increasing cement thickness. This indicates that it is valuable to keep as thin cement
thickness as possible. It is recognized that this stage of the process I critical. The cement material is
applied manually and painted on the dentin. This makes it sensitive as there is a human impact that
easily causes irregularities and flaws. Also the cement layer has to be prepared and applied within a
short working time, about 2 min [5]. The short timeframe can therefore also be a parameter causing
problem by the human factor.
Table 5 The maximum von Mises and Principal stresses in the different layers when varying the cement layer thickness.
Thickness [mm]
vM Dentin [MPa]
Princ. Dentin [MPa]
vM Cement [MPa]
Princ. Cement [MPa]
vM Ceramic
[MPa]
Princ. Ceramic
[MPa] 0.05 6.52 7.44 3.99 6.90 110 137 0.1 6.83 7.63 4.08 7.07 115 142 0.2 6.56 7.51 4.12 6.93 116 143 0.3 6.97 7.83 4.14 6.81 123 151 0.4 7.11 7.65 4.36 7.08 126 157
All maximum stresses occur somewhere at the border of the applied load area. It is seen from the
results above that the stresses increase almost exclusively for an increasing cement thickness
everywhere but in some of the maximum stresses for dentin and one of cement. This seems reliable
as it is probably due to that the cement material properties are more flexural and a thicker cement
layer allows a larger deformation in the ceramic layer which causes higher stresses. There is no clear
explanation to why some data seem to detach from this trend other than that the element division in
the model with 0.1 mm and 0.3 mm cement thickness is somewhat not perfectly identical to the
other models at these extreme values. Anyhow this fact does not cause a problem to the overall
analysis. The ceramic layer will be closest to failure independently on the maximum stresses in
underlying layers. As seen in Table 6 the ceramic is exclusively closest to failure.
Table 6 The percentage to failure when referring to bending and compression for all material layers.
Thickness [mm]
To failure in bending
Dentin [%]
To failure in compression Dentin [%]
To failure in bending
Cement [%]
To failure in compression Cement [%]
To failure in bending
Ceramic [%]
To failure in compression Ceramic [%]
0.05 3.1 2.5 11.7 5.1 29.4 38.1 0.1 3.2 2.6 12.0 5.2 30.7 39.4 0.2 3.1 2.5 12.1 5.1 31.0 39.7 0.3 3.3 2.6 12.2 5.0 32.9 41.9 0.4 3.4 2.6 12.8 5.2 33.7 43.6
11
Resulting stresses for simplified models The simplified models lead to an underestimation of the maximum stresses compared to the
anatomical model. This is shown in Figure 13. The diagram to the left shows the dentin and cement
layer and the diagram to the right shows the results for ceramic. These diagrams show that an
angled setup is more like the anatomical setup than the flat one.
Figure 13 Comparing the simplified and anatomic models in maximum stresses for the different layers.
The underestimation is large for all material layers and is probably not recommended for making
reliable simulations. It can be discussed that simplified modelling can be of value for comparable
simulation between the material layers. These results were somehow expected as the geometry is
much idealized in the simplified models.
Table 7 The expected failure loads for the different models
Failure criteria Failure load;
Flat model[N] Failure load;
Angled model [N] Failure load;
Anatomic model [N] von Mises 2920 1909 977 Rankine 2147 1616 759
The setup of the simplified flat model can be compared to previous testing [1]. The measured failure
load is in those experiments about 800 N. Considering that the geometry used in testing was oval
instead of circular and the E-modulus for the ceramic was 65 GPa in the experiments and here it was
94.4 GPa. This shows that the simulation might cause exceeding overestimations of crown strength
for such simple geometry when comparing to the first column in Table 7.
12
Different materials in Ceramic and Cement The ceramic and cement material were varied according to Table 2 and Table 3 respectively.
Varying Ceramic material
The ceramic was varied in order to see the crown material impact on the cement layer. The stress
results for the cement are seen as solid filled columns in Figure 14. The results show that the stresses
vary 140 % in bending and 100 % in compression when comparing highest to lowest maximum
stresses. It is seen from the transparent column in Figure 14 that the expected failure is far from
reaching for all different ceramic layers. The expected failure load is calculated from the assumption
of linear behaviour between applied load and resulting stresses. The stress at failure is the flexural
strength of the cement material and from this an expected failure load is calculated. The
Experimental Ceramic and the Hypothetical Ceramic are the only ceramics that imply an expected
failure load of about 1000 N and could indicate that the cement might fail before the ceramic crown.
The failure load for the ceramic is mentioned in Table 7 for the anatomic model. This seems a bit odd
as the elastic moduli at the ceramic are very different. This could be an indication that the elastic
modulus in combination of Poisson’s ratio is important. Also there is a possibility that the maximum
stresses do not occur at the exact same node for the different material setups. These are not real
materials but it shows that the materials available on the market probably have relatively good
properties to keep the cement layer from failing.
Figure 14 The maximum stresses in cement layer when varying the ceramic material are seen by solid filling. The transparent columns show how long is remained until cement reaches failure.
13
Varying Cement material
Similarly the diagram in Figure 15 shows how the cement layer properties affect the maximum
stresses in the ceramic crown. The resulting maximum stresses in both bending and compression
varies with about 7 % at most comparing the highest to lowest maximum stress. This implies that the
cement layer properties have little effect on the ceramic crown strength and more care should be
taken when choosing ceramic material than cement material.
Figure 15 The maximum stresses in ceramic layer when varying the cement material are seen by solid filling. The transparent columns show how long is remained until ceramic reaches failure.
In the same study as mentioned before in the section “Resulting stresses for simplified models”,
testing was made of an anatomical setup too [1]. Three different cement layers were used, which
were also used for this study in simulation. The different failure loads and expected failure loads can
be seen in Table 8, without mentioning the standard deviation of the test results.
14
Table 8 Comparing the expected failure load from simulations with test study failure load.
Cement material Failure load
comparison study [1] [N]
Expected Failure load; vM criteria [N]
Expected Failure load; Rankine criteria [N]
Panavia F 2.0, dual-polymerized phosphate-modified resin cement
550 1000 780
Rely X Unicem, dual-polymerized self-adhesive universal resin cement
400 990 770
Super bond C&B, self-cure dental adhesive resin cement
650 940 730
From Table 8 it is seen that generally the simulations seem to overestimate the strength of the
crown. When looking into detail it seems to be none at all similarities between the comparisons of
cement materials. Panavia F 2.0 seems to be the strongest cement material according to simulations
and Super bond C&B seems to be the strongest according to testing. The high diversity of these
results may say something about the importance of correct geometries. In the simulations the
geometries are always kept perfectly constant comparing to testing where flaws easily occur.
15
Conclusions The aim of the project was to show that loading area and position has a great impact on the stress
result in a dental crown. This was achieved by the results provided and all the most important
conclusions are listed below.
Loading area and the position of loading seems to be a very important factor for the
maximum stresses and indirectly the apparent strength of the dental crown.
A standardized testing method should be established in order to make fair comparisons
between testing.
More care has to be taken when choosing ceramic material than cement material in order to
control the crown strength.
A thick cement layer indicates to have a negative influence on the crown strength.
The simplified model seems to lead to a risky overestimation of the crown strength.
However, these simplified models can have a value in qualitatively comparing and ranking
materials while the large geometry effect on strength is rumored.
Practical testing seems to be very sensitive to flaws, load application and alignment.
16
References
[1] S. Shahrbafa, R. van Noorta, N. Martin, E. Ghassemieh and B. Mirzakouchakib, “Fracture strength of machined ceramic crowns as afunction of tooth preparation design and the elasticmodulus of the cement,” Dental Materials, pp. 234-241, 2013.
[2] J. R. Kelly, “Clinically relevant approach to failure testing of all-ceramic restorations,” 1999.
[3] T. Sornsuwan, A. Ellakwa and M. V. Swain, “Occlusal geometrical considerations in all-ceramic pre-molar crown failure testing,” Dental Materials, vol. 27, p. 1127–1134, 2011.
[4] D. Zhang, C. Lu, X. Zhang, S. Maob and D. Arola, “Contact fracture of full-ceramic crowns subjected to occlusal loads,” Journal of Biomechanics, vol. 41, p. 2995–3001, 2008.
[5] 3M ESPE, “RelyX™ Unicem; Self-Adhesive Universal Resin Cement; Technical Product Profile,” 2002. [Online].
[6] J. Kinney, S. Marshall and G. Marshall, “The Mechanical Properties of Human Dentin: a Critical Review and Re-evaluation of the Dental Literature,” Critical Reviews in Oral Biology and Medicine, vol. 14, no. 1, pp. 13-29, 2003.
[7] Michal Staninec,Grayson W. Marshall,Joan F. Hilton,David H. Pashley,Stuart A. Gansky,Sally J. Marshall and John H. Kinney, “Ultimate tensile strength of dentin: Evidence for a damage mechanics approach to dentin failure,” Journal of Biomedical Materials Research, vol. 63, no. 3, p. 342–345, 2002.
[8] S. Shahrbafa, R. van Noorta, B. Mirzakouchakib and E. Ghassemieh, “Effect of the crown design and interface lute parameters on the stress-state of a machined crown–tooth system: A finite element analysis,” Dental Materials, p. 9, 2013.
[9] B. Dejaka, A. Mlotkowskib and C. Langota, “Three-dimensional finite element analysis of molars with thin-walled prosthetic crowns made of various materials,” Dental materials, vol. 28, pp. 433-441, 2012.
[10] 3D-CoForm, “MeshLab,” 2012.
[11] D. S. S. Corp., “Solidworks”.
[12] “Solid Edge,” ST5.
[13] “ANSYS,” 14.5.
[14] P. David H. Johnson, “http://www.ansys.com,” 2002. [Online]. Available: http://www.ansys.com/staticassets/ANSYS/staticassets/resourcelibrary/confpaper/2002-Int-ANSYS-Conf-201.PDF. [Accessed 20 02 2013].
[15] J. Buller, “Comparison of structural integrity of dental crown systems of CAD/CAM and pressed,” Queen's University, Belfast, 2014.